Physics-Informed Few-Shot Flank-Wear Prediction from Sparse Inspections in Swiss-Type Turning of Stainless Steel

Flank-wear prediction for a new tool usually relies on few microscope inspections. This study tests how far sparse inspections extrapolate in Swiss-type turning of 1.4571 stainless steel with coated-carbide inserts. A 3 × 3 × 2 factorial design (cutting speed, feed, cooling) was run once per condition: fourteen short tests of five parts, with maximum flank wear (VBmax) measured on alternate parts, and four long tests run to tool breakage, giving 104 VBmax values. Models were evaluated by leave-one-tool-out (LOTO) validation. The vibration-feature regressors tested did not transfer to unseen tools (R2 between −1.06 and −0.11, worse than predicting the mean). From three inspections, a physics-informed power-law wear model (monotone, decelerating) with a fleet-shared exponent extrapolated the long tests (VBmax ≤ 300 µm) with a mean absolute error of 16 µm, against 41 µm for a fleet-average curve and 33 µm for linear extrapolation; fixed-exponent and logarithmic fits gave 20–27 µm, so the concave shape carried most of the gain. With two inspections, the law offered no advantage. Finite LOTO conformal bands reached 90–95% empirical coverage (target 90%) but were 76–130 µm wide. Four breakages, after last inspections of 154–279 µm, were too few to identify a wear-based stop threshold.

Authors

Institutions

Publication Details

Journal
Machines
Published
2026-11-19
DOI
https://doi.org/10.3390/machines14111378
Primary Topic
Advanced machining processes and optimization
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Physics-Informed Few-Shot Flank-Wear Prediction from Sparse Inspections in Swiss-Type Turning of Stainless Steel

Saman Fattahi, Josef Schnell, Bahman Azarhoushang, Seyyed Sajjad Ahmadpoor et al.
Machines
Advanced machining processes and optimization
article

Physics-Informed Few-Shot Flank-Wear Prediction from Sparse Inspections in Swiss-Type Turning of Stainless Steel

Saman Fattahi, Josef Schnell, Bahman Azarhoushang, Seyyed Sajjad Ahmadpoor, Amir Alinaghizadeh, Manuel Inocencio Pusma Jurado
article en

Abstract

Flank-wear prediction for a new tool usually relies on few microscope inspections. This study tests how far sparse inspections extrapolate in Swiss-type turning of 1.4571 stainless steel with coated-carbide inserts. A 3 × 3 × 2 factorial design (cutting speed, feed, cooling) was run once per condition: fourteen short tests of five parts, with maximum flank wear (VBmax) measured on alternate parts, and four long tests run to tool breakage, giving 104 VBmax values. Models were evaluated by leave-one-tool-out (LOTO) validation. The vibration-feature regressors tested did not transfer to unseen tools (R2 between −1.06 and −0.11, worse than predicting the mean). From three inspections, a physics-informed power-law wear model (monotone, decelerating) with a fleet-shared exponent extrapolated the long tests (VBmax ≤ 300 µm) with a mean absolute error of 16 µm, against 41 µm for a fleet-average curve and 33 µm for linear extrapolation; fixed-exponent and logarithmic fits gave 20–27 µm, so the concave shape carried most of the gain. With two inspections, the law offered no advantage. Finite LOTO conformal bands reached 90–95% empirical coverage (target 90%) but were 76–130 µm wide. Four breakages, after last inspections of 154–279 µm, were too few to identify a wear-based stop threshold.

MachinesVol. 14(11)
University of Freiburg (DE), Islamic Azad University Semnan (IR), National University of Engineering (PE), Furtwangen University (DE)
Advanced machining processes and optimization
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon